Creating a tool for Exhaustive Code Coverage Test Case Generation Using ML/DL

I want to create/use a tool that can automatically generate test cases for a Java Program/Package which should be able to provide exhaustive coverage. Is there a way the test cases can be generated using Deep Learning/Machine Learning? The inputs for the program that I want to run it for consist of discrete as well as continuous values and even Strings(can be converted to discrete values). The OUTPUT should be a test case that increases the coverage of the package after getting appended to the existing package. The CONVERGENCE CRITERIA can be the Total Coverage getting increased.